Lesson 04

Getting and Setting Series Values

Read Series values by label with loc and by position with iloc, slice both ways, pass lists of labels, and assign new values back.

Intermediate25 min

Getting and Setting Series Values hero Module 24 Lesson 4

A Series has two ways to reach a value: the label and the position. loc always means the label and iloc always means the position. Being explicit avoids the classic confusion when the labels are themselves numbers.

A labelled Series with the loc and iloc targets highlighted

By label and by position

loc takes a label and raises KeyError when it is missing. iloc takes a position, counts from 0, accepts negatives, and raises IndexError when it runs past the end.

python

import pandas as pd

names = pd.Series(["Avery", "Jordan", "Casey", "Riley", "Morgan"],
                  index=["a", "b", "c", "d", "e"])
print(names.loc["d"])
print(names.iloc[0], names.iloc[-1])

try:
    names.loc["k"]
except KeyError as error:
    print("KeyError:", error)

Output:

text

Riley
Avery Morgan
KeyError: 'k'

Slices differ at the end

A position slice leaves the stop out, exactly like a list. A label slice keeps the stop label in, because pandas cannot know what comes after it. This is the one difference people trip over.

python

import pandas as pd

names = pd.Series(["Avery", "Jordan", "Casey", "Riley", "Morgan"],
                  index=["a", "b", "c", "d", "e"])
print(names.iloc[1:4])
print(names.loc["b":"e"])

Output:

text

b    Jordan
c     Casey
d     Riley
dtype: object
b    Jordan
c     Casey
d     Riley
e    Morgan
dtype: object

Several values at once

Pass a list of labels to loc, or a list of positions to iloc. The result is a new Series in the order you asked for, and a label may repeat.

python

import pandas as pd

names = pd.Series(["Avery", "Jordan", "Casey", "Riley", "Morgan"],
                  index=["a", "b", "c", "d", "e"])
print(names.loc[["d", "e", "a"]])
print(names.iloc[[2, 3, 0]])

Output:

text

d     Riley
e    Morgan
a     Avery
dtype: object
c    Casey
d    Riley
a    Avery
dtype: object

Change values

The same accessors work on the left of =. Assign one value, or assign to a slice or a list of labels to set several at once. A Series is mutable, so this changes it in place.

python

import pandas as pd

names = pd.Series(["Avery", "Jordan", "Casey", "Riley", "Morgan"],
                  index=["a", "b", "c", "d", "e"])
names.iloc[2] = "Quinn"
names.loc["e"] = "Harper"
names.loc["a":"c"] = "unassigned"
print(names)

Output:

text

a    unassigned
b    unassigned
c    unassigned
d         Riley
e        Harper
dtype: object

Add a label

Assigning to a label that does not exist appends it. That is the difference between a Series and a Python list, where an unknown position is an error.

python

import pandas as pd

scores = pd.Series({"Avery": 88, "Jordan": 74})
scores.loc["Casey"] = 91
print(scores)
print(scores.get("Riley", "not found"))

Output:

text

Avery     88
Jordan    74
Casey     91
dtype: int64
not found

Key takeaways

loc is labels, iloc is positions
iloc[1:4] leaves the stop out. loc['b':'e'] keeps it in
A list inside either one returns several values, in your order
Assigning to a new label appends it
Next: Series Statistics
Previous: Series Attributes

Loading playground…

Check your understanding

Getting and Setting Series Values Quiz

What does loc use?

1 / 10

← Series Attributes

Series Statistics →